Focus fusion and other plasma focus devices are inertial confinement devices. Inertial confinement relies on the momentum of the matter to hold the fusing system together. PFs use a magnetic field to accelerate and try to confine the matter but in the end the field cannot hope to win. One way to look at the problem is with pressure. The magnetic field has a pressure associated with it. As the current increases the magnetic field increases and as the inner radius of the plasma column decreases, the magnetic field increases. Magnetic pressure increases with magnetic flux squared. The kinetic pressure is your normal density times Boltzman constant times temperature. When you strike the plasma, the magnetic pressure overwhelms the kinetic pressure and pushes the mass forward just like pushing a door shut. When plasma reaches the end of the electrodes, the magnetic field can press the plasma into a cylindrical object commonly called a pinch. Within this pinch region hot spots which are sometimes called plasmoids and other regions form due to instabilities in the compression process. One average, the density increases and kinetic energy of the moving plasma is converted to thermal energy as the pinch stagnates. You can go in any radial direction as the magnetic pressure is pushing you toward the center. The stagnation is caused when magnetic pressure and kinetic pressure equalize. Before stagnation, the current starts to drop sharply due to electrical changes in the circuit as a result of changing the plasma shape. Thus, kinetic pressure eventually overwhelms the magnetic pressure and the pinch explodes. To increase fusion time you must hold up the current, which has not proven to be positive. You want the rapid drop in current. If you push more mass but the plasma is colder and less likely to reach the conditions of interest. The fusion lifetime for a PF device appears to be fixed at ~10 ns.
What I was thinking was to variate the current. Like decrease or increase or during the cycle, to just keep it going a little bit longer. But that maybe to tricky when you just have nanoseconds in each cycle.
You can try and fire switches at different times but the jitter in the switches tends to be too great to control. You also run into issues with different voltages in the circuit at different times. They are pulse shaping systems that fire switches at different times but they cannot respond at the nanosecond scale at the MA level.